cephfs-client Segmentation fault with not-root mount point

yu2xiangyang <yu2xiangyang@xxxxxxx> · Sun, 18 Sep 2016 10:14:29 +0800 (CST)

My envirenment is displayed below:
The ceph-fuse client is 10.2.2, and the ceph osd is 0.94.3,
details beblow:

[root@localhost ~]# rpm -qa | grep ceph
libcephfs1-10.2.2-0.el7.centos.x86_64
python-cephfs-10.2.2-0.el7.centos.x86_64
ceph-common-0.94.3-0.el7.x86_64
ceph-fuse-10.2.2-0.el7.centos.x86_64
ceph-0.94.3-0.el7.x86_64
ceph-mds-10.2.2-0.el7.centos.x86_64

 [root@localhost ~]# rpm -qa | grep rados
librados2-devel-0.94.3-0.el7.x86_64
librados2-0.94.3-0.el7.x86_64
libradosstriper1-0.94.3-0.el7.x86_64
python-rados-0.94.3-0.el7.x86_64

 When I mount the cephfs with "ceph-fuse -m  10.222.5.229:6789  --client-mount /client_one /mnt/test",I got ceph-client crash when I run a few hours later.

-16> 2016-08-18 18:37:54.134672
7fd552ffd700  3 client.214296 ll_flush
0x7fd5307e8520 10000478575 

-15> 2016-08-18 18:37:54.134717 7fd5128e2700  3 client.214296 ll_release (fh)0x7fd5307e8520
10000478575 

-14> 2016-08-18 18:37:54.134725 7fd5128e2700  5 client.214296 _release_fh 0x7fd5307e8520
mode 1 on 10000478575.head(faked_ino=0 ref=3 ll_ref=11030
cap_refs={1024=0,2048=0} open={1=1} mode=100644 size=12401/0 mtime=2016-08-17
13:49:59.382502 caps=pAsLsXsFscr(0=pAsLsXsFscr) objectset[10000478575 ts 0/0 objects
1 dirty_or_tx 0] parents=0x7fd55c0120d0 0x7fd55c011b30)

-13> 2016-08-18 18:37:54.136109 7fd551ffb700  3 client.214296 ll_getattr 1000047417f.head

-12> 2016-08-18 18:37:54.136118 7fd551ffb700  3 client.214296 ll_getattr 1000047417f.head =
0

   -11>
2016-08-18 18:37:54.136126 7fd551ffb700 
3 client.214296 ll_forget 1000047417f 1

-10> 2016-08-18 18:37:54.136133 7fd551ffb700  3 client.214296 ll_lookup 0x7fd55c0108d0 2016

-9> 2016-08-18 18:37:54.136140 7fd551ffb700  3 client.214296 ll_lookup 0x7fd55c0108d0 2016
-> 0 (10000474182)

-8> 2016-08-18 18:37:54.136148 7fd551ffb700  3 client.214296 ll_forget 1000047417f 1

-7> 2016-08-18 18:37:54.136181 7fd5527fc700  3 client.214296 ll_getattr 10000474182.head

-6> 2016-08-18 18:37:54.136189 7fd5527fc700  3 client.214296 ll_getattr 10000474182.head =
0

-5> 2016-08-18 18:37:54.136735 7fd550c92700  2 -- 10.155.2.5:0/1557134465 >>
10.155.2.5:6820/4511 pipe(0x7fd54c012ef0 sd=2 :48226 s=2 pgs=107 cs=1 l=1
c=0x7fd54c0141b0).reader couldn't read tag, (0) Success

-4> 2016-08-18 18:37:54.136792 7fd550c92700  2 -- 10.155.2.5:0/1557134465 >>
10.155.2.5:6820/4511 pipe(0x7fd54c012ef0 sd=2 :48226 s=2 pgs=107 cs=1 l=1
c=0x7fd54c0141b0).fault (0) Success

-3> 2016-08-18 18:37:54.136950 7fd56bff7700  1 client.214296.objecter ms_handle_reset on
osd.5

-2> 2016-08-18 18:37:54.136967 7fd56bff7700  1 -- 10.155.2.5:0/1557134465 mark_down
0x7fd54c0141b0 -- pipe dne

-1> 2016-08-18 18:37:54.137054 7fd56bff7700  1 -- 10.155.2.5:0/1557134465 -->
10.155.2.5:6820/4511 -- osd_op(client.214296.0:630732 4.a8ddcaa5
10000493bde.00000000 [write 0~12401] snapc 1=[] RETRY=1
ondisk+retry+write+known_if_redirected e836) v7 -- ?+0 0x7fd55ca2ff40 con
0x7fd55ca6d710

     0> 2016-08-18
18:37:54.141233 7fd5527fc700 -1 *** Caught signal (Segmentation fault) **

 in thread 7fd5527fc700 thread_name:ceph-fuse

 ceph
version 10.2.2 (45107e21c568dd033c2f0a3107dec8f0b0e58374)

 1:
(()+0x29eeda) [0x7fd57878feda]

 2:
(()+0xf130) [0x7fd577505130]

 3:
(Client::get_root_ino()+0x10) [0x7fd57868be60]

 4:
(CephFuse::Handle::make_fake_ino(inodeno_t, snapid_t)+0x18d) [0x7fd57868992d]

 5:
(()+0x199261) [0x7fd57868a261]

 6:
(()+0x164b5) [0x7fd5780a64b5]

 7:
(()+0x16bdb) [0x7fd5780a6bdb]

 8:
(()+0x13471) [0x7fd5780a3471]

 9:
(()+0x7df5) [0x7fd5774fddf5]

 10:
(clone()+0x6d) [0x7fd5763e61ad]

But when I mount the cephfs with "ceph-fuse -m 10.222.5.229:6789   /mnt/test", ceph fuse client runs all right during few days.
I do not think the problem is related with the 0.94.3 OSD.
Does someone encouter the same problem with cephfs10.2.2?

 _______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com